Install and load tidyverse, sf and Distance packages
Unzip the data folder and place in your root directory
Objective
Today we will be estimating the total size of a population of blue whales using data from line-transect surveys.
Photo: Russell Leaper
Instructions
We will be using data from a simulated line-transect survey for blue whales off the coast of California, Oregon, and Washington.
1) Read in your data (simulated_bw.csv) and look at it. What kind of information is in each of the columns?
2) Plot the locations of each detection on top of the tracklines. You will need to read in the tracklines.shp and region.shp files using read_sf(). You can plot these using geom_sf().
3) Make a histogram of your detection distances.
4) Run three distance sampling models using the ds() function in the Distance package. You will need to select a detection function for each model: half-normal = "hn", hazard rate = "hr", and uniform = "unif". You will also need to set convert_units = convert_units("meter", "meter", "meter") inside of your function call. Look at the summary function for each model and plot the output.
Instructions part 2
5) Use gof_ds() to assess whether your model fits the data.
6) Compare the three models using AIC(). Which of these is the best model?
7) Extract your estimated abundance and 95% confidence intervals from each model. Do the confidence intervals contain our “true” estimate of 800 individuals?
5) Use gof_ds() to assess whether your model fits the data.
6) Compare the three models using AIC(). Which of these is the best model?
7) Extract your estimated abundance and 95% confidence intervals from each model. Do the confidence intervals contain our “true” estimate of 800 individuals?
Discussion
Which model is the best model?
What might be some challenges with fitting this kind of model? How can we account for these?
